In this paper we study pose estimation for general non-central cameras, using planar targets. The method proposed uses non-minimal data. Using the homography matrix to represent the transformation between the world and camera coordinate systems, we describe a non-iterative algorithm for pose estimation. To improve the accuracy of the solutions, data-set normalization is used. In addition, we propose a parameter optimization to refine the pose estimate. We evaluate the proposed solutions against the state-of-the-art method (for general targets) in terms of both robustness to noise and computation time. From the experiments, we show that the proposed method plus normalization is more accurate against noise and less sensitive to variations of the imaging device. We also show that the numerical results obtained with this method improve with the increasing number of data points. In terms of processing speed, the versions of the algorithm presented are significantly faster than the state-of-the-art algorithm. To further evaluate our method, we performed an experiment of a simple augmented reality application in which we show that our method can be easily applied.